Porous wood decorated with gold nanoparticles as flow-through membrane reactor for catalytic hydrogenation of methylene blue and 4-nitrophenol

Wood has a natural three-dimensional porous structure. As one main component of wood, lignin is rich in a variety of reducing functional groups which can in situ reduce Au (III) into Au (0). Through this efficient, green and convenient method, gold nanoparticles (Au NPs) were generated and anchored in Paulownia Sieb. et Zucc. Chip to fabricate an Au NP/Wood membrane reactor. The characterization of Au NP/Wood by SEM, XRD and XPS showed that Au NPs were uniformly dispersed in wood. The catalytic capacity of the reactor for the reduction of methylene blue (MB) and 4-nitrophenol (4-NP) were evaluated. The Au NP/Wood exhibited the substantial catalytic capacity and the reaction rate constants were 0.162 min−1 and 0.152 min−1, respectively. As the filter membrane, the flowing catalytic capacity investigation revealed that the hydrogenation of MB and 4-NP was over 98% as the flux was 0.973 × 103 L/m2·h. Even after eight cycles, the catalytic capacity of the membrane slightly decreased, while the hydrogenation still remained above 90%. This green synthetic Au NP/Wood has proved to be a viable and potential material for the treatment of dyes and nitroaromatic pollutants, in addition, using wood as a feedstock provides a sustainable feature of this work.

The Au NPs/Wood placed on the sand core filter was used for the continuous flow 145 degradation which was conducted at 0.015 MPa. Before the reaction, the Au NPs/Wood 146 was immersed in the MB solution to reach the equilibrium to eliminate the adsorption 147 interference (Liu et al., 2017). Thereafter, the mixture of MB and NaBH4 was filtered 148 through the Au NPs/Wood to complete the continuous flow degradation. The flow 149 degradation of 4-NP was conducted with the same procedure. As a comparison, the 150 wood chip was also used to conduct the continuous flow degradation. The factors 151 including the Au NPs/Wood layers, the concentration of contaminants, and the water 152 flux were evaluated. As for the evaluation of recyclability, the Au NPs/Wood was 153 recovered by washing with deionized water followed with ethanol for the subsequent 154 reaction. 155

Characterizations 157
Under the sunlight irradiation, Au (III) was diffused into the wood and reduced in 7 situ by lignin to form Au NPs. Fig. 1a shows that the wood color changed from brown 159 yellow to purple due to the surface plasmon effect of Au NPs depositing on the surface 160 of wood cell wall (Amendola et al., 2017). Fig. 1b indicates that Au NPs was uniformly 161 deposited throughout the wood slice. SEM-EDS characterization observed that the 162 wood cell wall surface was uniformly loaded with Au NPs (Fig. 1c-f). The high 163 resolution SEM image (Fig. 1g) confirmed that Au NPs immobilized uniformly on the 164 wood cell wall surface without aggregation which retained their catalytic ability. 165 Furthermore, the gold content of the sample was quantified as 1.4wt% by ICP-MS. The 166 unique three-dimensional porous structure was maintained in the Au NPs/Wood. A large 167 number of honeycomb like units and channels were connected with the perforated 168 plates containing nano size pores (Fig. 1h). Fig. 1i confirms that nano size pores were 169 identified in the perforated plates. Fig. 1j demonstrates that the vessels connected with 170 the perforated plates can be used as channels for transferring water and nutrients in the 171 wood. All of these characteristics of the wood ensured that the Au NPs/Wood can be 172 used as an ideal catalytic membrane for wastewater treatment.

Batchwise catalytic degradation 205
The UV-Vis absorbance at 665 nm decreased slightly with the reaction time as the 206 wood slice was immersed in the mixture (Fig. 3a), which was attributed to a little 207 adsorption of MB by the wood. As a comparison, the absorbance at 665 nm was little 208 detected after 21 min immersion (Fig. 3b), indicating that the Au NPs/Wood had an 209 effective catalytic performance on the reduction of MB. According to the previous 210 report, the reduction process can be described as a quasi-first-order kinetic process 211 (Ramirez et al., 2017). Thus, the reaction rate constant (k) can calculated to evaluate 212 the catalyst's performance according to the following equation: 213 Where 0 and are the concentration of the contaminant at the initial stage and at 215 the reaction time t (min), respectively, 0 and are the corresponding absorbance, 216 k is the reaction rate constant. 217 Fig. 3c shows the quasi-first-order kinetics between the reactants, i.e. the reaction 218 rate constant k was calculated from the linear relationship between ln(Ct/C0) and 219 reaction time. The catalytic reaction rate k of wood chips for MB and NaBH4 was 220 calculated to be 0.008 min −1 , and the k value for MB catalytic reduction using the Au 221 NPs/Wood was calculated to be 0.162 min -1 . It can be seen from Table 1 that the 222 10 catalytic activity of Au NPS /Wood is comparable to some extent, or even higher than 223 that of other materials supported gold nanoparticles in the catalytic reaction kinetics of 224

MB. 225
As for the catalytic reduction of 4-NP, the characteristic absorbance at the 226 wavelength of 400 nm displayed by UV-Vis decreased slightly with the reaction time 227 (Fig. 3d), which was attributed to the small amount of adsorption of 4-NP by wood. 228 When Au NPs/Wood was applied to the pollutants, the UV-Vis scanning spectrum 229 showed that the characteristic peak at 400nm almost disappeared after 27min reaction, 230 while a new peak appeared at 298nm, which was the characteristic peak of 4-231 aminophenol, indicating that 4-NP was reduced to 4-AP ( Au NPs/Wood for mixture were calculated as 0.011 min -1 and 0.152 min -1 , respectively 235 ( Fig. 3f). It can be seen from Table 1

244
The Au NPs were isolated from the Au NPs/Wood after 2 h ultrasound treatment 245 at room temperature, and the characteristic color of Au NPs was observed from the 246 solution (Unal et al., 2020) (Fig. S2a). As the suspension had stood for 2 h, the 247 precipitant was observed at the bottom of the centrifuge tube, indicating that the solo 248 Au NPs was inclined to precipitate (Fig. S2b). Therefore, the catalytic performance of 249 12 Au NPs was attenuated by the precipitation. For example, the UV absorbance in the 250 range of 400-800 nm decreased little even after the mixture of MB and NaBH4 had 251 mixed with Au NPs for 19 min (Fig. S2c). This result verified that the decoration of Au 252 NPs on the surface of wood cell wall prevented their precipitation, thereby retained 253 their catalytic performance. 254

Continuous flow catalytic degradation 255
According to the phenomena and UV-Vis measurement results observed in Fig. S2,  256 it can be inferred that the Au NPs/Wood has a promising application in continuous flow 257 catalytic reaction. Taking MB as a model pollutant, the flow catalytic ability of Au 258 NPs/Wood was evaluated under the vacuum filtration device shown in Fig. S4a. It can 259 be clearly seen that the initial blue solution becomes colorless after the filtration (Fig.  260 S4b). During the filtration process, the solution contacted the Au NPs in the wood, 261 therefore, the MB was catalytically reduced to LMB. When the vacuum was maintained 262 as 0.015 MPa, the impact of different layer of Au NPs/Wood on the reduction of MB 263 (30 mg/L) was evaluated. The results showed that the catalytic degradation efficiency 264 of the one-layer membrane was only 22%. As more layers of Au NPs/Wood were 265 applied, the removal efficiency of pollutants and the filtration time were increased. For 266 example, the removal efficiency was obtained as 98.9% as three layers of Au NPs/Wood 267 were applied, however the filtration time increased to 12 min (Fig. 4a). This result was 268 explained by the increase of the axial tracheid perforation plate, which retained the 269 mixture in the wood cavity, slowing down the flow velocity by forcing the solution flow 270 through more the perforated plates, and increased the contact between MB and Au NPs. 271 Therefore, the membrane composing with three layers of Au NPs/Wood were selected 272 for the next flow catalytic experiments (Thomas, 1977). 273 The impact of water fluxes on the catalytic efficiency of the Au NPs/Wood was 274 evaluated, as well. Technically, the water flux was adjusted by the vacuum. The flux of 275 water treatment can be calculated by following equation: 276 Where V is the volume of solution (L), T is the required time for a specific volume 278 13 solution filtering through the membrane (h), A is the effective area of the membrane 279 (m 2 ), and J is the treatment flux (L/m 2 ·h) (Cheng et al., 2018). 280 Even the flux was increased to 0.973×10 3 L/m 2 ·h (the corresponding vacuum was 281 of 0.015 MPa), the removal efficiency was still more than 95% (Fig. 4b). This result 282 was attributed to three explanations. Firstly, Au NPs, as the electron transfer between 283 nucleophilic NaBH4 and electrophilic contaminant, distribute uniformly in wood, 284 which promotes the catalytic reduction process. Secondly, the porous structure and the 285 perforated plates of the wood retain the effective contact between catalyst and reagent. 286  The absorbance change during the filtration was analyzed by UV-Vis in the range of 305 200-600 nm. As shown in Fig. 5a, the peak strength at 400 nm after the wood filtration 306 hardly decreased, which can be interpreted as the finite adsorption of 4-NP by the wood. 307 However, After the Au NPs/Wood treatment, the characteristic peak at 400 nm almost 308 disappeared, meanwhile the absorption peak of 4-AP at 298 nm appeared, indicating 309 that 4-NP was effectively catalytically reduced to 4-AP (Fig. 5b).

Recyclability of Au NPs/Wood 314
The recyclability is another important parameter of polyphase catalyst for the 315 practical application. In order to investigate the recyclability of Au NPs/Wood, it was 316 recycled eleven times for the catalytic reduction of MB and recycled eight times for 4-317 NP. Fig. 6a shows that after 11 times recycle, the removal efficiency of MB by the Au 318 NPs/Wood was still over 85%. The catalytic reduction efficiency of 4-NP by the Au 319 NPs/Wood remained above 85% after 8 cycles (Fig. 6b). Therefore, the Au NPs/Wood 320 had robust stability and reusability, which has important guiding significance for its 321 application in practical wastewater treatment. We speculated that there are three main 322 reasons for the difference recyclability of Au NPs/Wood for the degradation of MB and 323 4-NP. Firstly, MB contains unsaturated bond in its molecular formula, and its induction 324 effect is stronger than that of 4-nitrophenol (